Membrane Glycoprotein, to Escherichia coli Expressing - Infection and ...

6 downloads 93 Views 2MB Size Report
E. coli HB101, which does not express fimbriae (Fim-) (8, ..... fimbriae will be examined in other studies. ..... 49. von Kleist, S., G. Chavanel, and P. Burtin. 1972.
INFECTION AND IMMUNITY, JUlY 1991, p. 2485-2493 0019-9567/91/072485-09$02.00/0 Copyright ©3 1991, American Society for Microbiology

Vol. 59, No. 7

Binding of Nonspecific Cross-Reacting Antigen, a Granulocyte Membrane Glycoprotein, to Escherichia coli Expressing Type 1 Fimbriae SYBILLE L. SAUTER,1 SHANE M. RUTHERFURD,1 CHRISTOPH WAGENER,2 JOHN E. SHIVELY,' AND STANLEY A. HEFTAl*

Division of Immunology, Beckman Research Institute of the City of Hope, Duarte, California 91010,1 and Department of Clinical Chemistry, University Hospital Eppendorf, 2000 Hamburg 20, Germany2 Received 30 October 1990/Accepted 3 May 1991

Nonspecific cross-reacting antigen (NCA) is a well-characterized membrane glycoprotein on granulocytes, macrophages, and lung epithelium. Structural studies at the protein and genomic levels have revealed that NCA is a member of the immunoglobulin supergene family, and hybridization studies showed that the transcript level of NCA is induced by treatment with gamma interferon. These studies, as well as the expression of NCA on granulocytes, suggest a role for NCA in immune response. For a first step in studying this possible role, we have examined the binding of two glycoforms of NCA designated NCA-50 (Mr, 50,000) and TEX-75 (M,, 75,000). Here we report the results from binding assays which demonstrate carbohydrate-mediated binding of Escherichia coli expressing type 1 fimbriae and of isolated type 1 fimbriae to NCA-50. TEX-75 did not bind to the purified fimbriae but bound slightly to the bacterial strain. Inhibition studies showed that the binding to NCA-50 involved interaction of mannose moieties on NCA-50 and lectins on the fimbriae. The binding of NCA-50 to bacterial fimbriae was confirmed by electron microscopy studies, using immunolabeling techniques. In addition, we show that the surface expression of NCA-50 (and presumably of other NCA species) on isolated polymorphonuclear leukocytes is increased following activation with the bacterial peptide formylmethionyl-leucyl-phenylalanine, consistent with a role for NCA in immune response. Nonspecific cross-reacting antigen (NCA) is a highly glycosylated membrane protein found on granulocytes, macrophages, and lung epithelium and in colonic adenocarcinoma. The first NCA form (50 kDa) (NCA-50) was identified by von Kleist et al. (49) and Mach and Pusztaszeri (31) in 1972 and named for its immunological cross-reactivity with carcinoembryonic antigen (CEA), a well-characterized tumor-associated antigen (43). Since then, several other NCA forms with apparent molecular masses of 50, 75 (designated TEX75, for tumor-extracted antigen), 90, 95, and 160 kDa (3, 9, 27, 31, 49) have been identified. Protein and nucleotide sequencing have resolved much of the confusion regarding the structures and relatedness of these antigens (2, 18, 21, 32, 35, 38, 46, 48). NCA-50 and TEX-75 were found to have identical protein sequences and numbers of glycosylation sites but to be distinct glycoproteins differing significantly in carbohydrate content (21). These same studies revealed the likelihood that TEX-75 (isolated in this laboratory) (27) is equivalent to the NCA-90 identified by Audette et al. (3), at least with regard to the protein sequence. NCA-95 was found by others to be a unique antigen and to share >85% sequence homology with NCA-50 and TEX-75 (9). Tissue distribution studies have revealed that NCA-95 is found on granulocytes only, while NCA-50 and TEX-75 are found on granulocytic and epithelial cells (9). Although not confirmed, it is generally believed that the 160-kDa form of NCA is biliary glycoprotein, a structurally related (4, 22, 23), but distinct, glycoprotein. To date, only two genes are known for NCA, one encoding NCA-95 (2) and the other encoding NCA-50 or TEX-75 (32, 46), indicating that much of the size diversity observed for the NCA family members is due to

*

differences in the glycosylation patterns of the respective antigens. Although much is known concerning the structure and tissue distribution of these antigens, only recently have researchers reported possible biological roles for NCA and CEA. Oikawa et al. (34) proposed that NCA functions as a homotypic adhesion molecule, a function similar to that suggested for CEA (6). Oikawa's conclusions were based on the observation of a slight increase (-10%) in cell-cell aggregation of Chinese hamster ovary cells following transfection of NCA cDNA. The relevance of this finding as it relates to NCA antigens on granulocytes, however, remains to be determined. In other studies designed to examine alternative roles, Leusch and coworkers (30) recently described the bacterial binding properties of CEA and provided evidence that NCA had similar properties. The role of cell surface glycoproteins in mediating bacterial and viral binding is well recognized (40, 51). Many bacteria express long proteinaceous appendages, called fimbriae or pili, that contain lectins which enable the microorganism to adhere to surface glycoproteins in a carbohydratespecific manner. Oligosaccharide sequences identified as receptors for fimbriae include Manal-3Man for type 1 fimbriae, NeuAca2-3Gal for S fimbriae, and Galotl-4Gal for P fimbriae (5, 41). The most common and thoroughly characterized of these structures are mannose-specific type 1 fimbriae that are found in about 70% of Escherichia coli strains (37). Numerous studies have demonstrated the involvement of type 1 fimbriae in bacterial binding to epithelial and phagocytic cells, and there is a positive correlation between the degree of expression of type 1 fimbriae and the susceptibility to phagocytosis by granulocytes in serum-free medium (33, 42). On the basis of the above information and the finding that NCA-50 contains a high amount of high-

Corresponding author. 2485

2486

SAUTER ET AL.

mannose sugars (see Results), we chose the mannose-specific type 1 fimbriae for our studies as a possible ligand for NCA-50. Here we extend the initial observations of Leusch et al. (30) and provide direct evidence that NCA-50 binds specifically to type 1 fimbriae. We further examine the type 1 fimbriae binding properties of TEX-75, the expression of NCA on granulocytes, and the carbohydrate compositions of NCA-50 and TEX-75.

MATERIALS AND METHODS Bacteria. E. coli 38 is a clinical isolate from a patient with a urinary tract infection and was obtained from the Department of Infectious Diseases, City of Hope Medical Center. E. coli HB101, which does not express fimbriae (Fim-) (8, 28), was purchased from the American Type Culture Collection. Both strains were grown at 37°C under static conditions for 48 h in Luria-Bertani (LB) broth and serially diluted three times before use. These conditions favor the growth of cells with the Fim+ phenotype over cells with the Fim- phenotype (36). Bacteria were tested in a yeast agglutination assay for their ability to mediate mannose-specific binding. Bacteria were considered mannose sensitive when visible agglutination could be prevented by the addition of 20 mM a-methylmannoside (MMan) but not with ot-methylgalactoside (MGal) or at-methylglucoside (MGlu). Confirmation that E. coli 38 expressed type 1 fimbriae was accomplished by using anti-type 1 polyclonal antibodies which were generously provided by S. N. Abraham, Washington University, St. Louis, Mo. (data not shown). Isolation of type 1 fimbriae. Type 1 fimbriae isolated from E. coli CSH50 (39) were a generous gift from M. B. Goetz, Department of Infectious Diseases, VA Medical Center, University of California, Los Angeles. The purification procedure followed the method of Eshdat et al. (12), and the purity of the fimbriae was examined by electron microscopy, amino acid analysis, and immunoreactivity with anti-type 1

polyclonal antibodies. Isolation of NCA-50 and TEX-75. Both glycoproteins were isolated from a liver metastasis of a colon adenocarcinoma as previously described (21). The purity of the glycoproteins was verified by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blot (immunoblot) analysis with the anti-NCA-reactive monoclonal antibody T84.1E3 (50). This antibody was raised against CEA and cross-reacts with other CEA family members, including NCA-50 and TEX-75. Concentrations of glycoproteins and type 1 fimbriae were determined by amino acid analysis, and the mole amounts given refer to the protein moiety only. Carbohydrate analysis and endo H digestion of NCA-50 and TEX-75. Approximately 50 pmol of NCA-50 and 120 pmol of TEX-75 were used for carbohydrate analysis performed by established procedures (20). For determination of neutral sugars, the protein samples were placed in glass hydrolysis tubes, lyophilized to dryness, and resuspended in 15 ,ul of 2 M aqueous trifluoroacetic acid. The tubes were sparged of oxygen with helium and sealed, and the samples were hydrolyzed for 4 h at 100°C. A similar technique was used for the detection of amino sugars and sialic acid, except that the samples were hydrolyzed with 6 M HCI for 4 h at 100°C or with 0.1 M trifluoroacetic acid for 8 h at 50°C, respectively. Following hydrolysis, the samples were dried, resuspended in deionized water, and analyzed with a Bio-LC system (Dionex Corp.) equipped with an amperometric detector. For digestion with endoglycosidase H (endo H) (Boehringer Mannheim), the samples (-150 pmol each) were suspended

INFECT. IMMUN.

in 10 pul of 0.12% SDS (wt/vol), boiled for 3 min, diluted with 15 pul of 50 mM NaH2PO4 buffer at pH 6.0, and digested with 5 pLI of endo H (2 U/ml) for 48 h at 37°C. After digestion, the samples were loaded onto a 10% SDS-polyacrylamide gel, electrophoresed, and blotted onto nitrocellulose with a PolyBlot electrotransfer system (American Bionetics). The blots were sequentially incubated with the anti-NCA-reactive monoclonal antibody T84.1E3 (mouse ascitic fluid diluted 1:500 in phosphate-buffered saline [PBS]) and horseradish peroxidase-labeled goat anti-mouse antibody (1:500 dilution), and the specific protein bands were visualized by addition of the chromogenic substrate 4-chloro-1-naphthol. Bacterial agglutination assay. E. coli 38 was washed in saline (0.15 M NaCl) and 5 x 108 bacteria were incubated with 2 puM NCA-50 in 20 RI of saline for 20 min at room temperature. A 10-,u aliquot was heat fixed on a glass slide, stained with crystal violet (1%) for 1 min, washed with deionized water, and examined under a light microscope. The glycoproteins TEX-75, ovalbumin, thyroglobulin, ribonuclease B, and fetuin and Fim- E. coli HB101 cells served as controls. Biotinylation of bacteria and type 1 fimbriae. Bacteria were washed twice in PBS and 50 mM bicarbonate buffer (pH 8.5), respectively. The cells were resuspended in the bicarbonate buffer containing 10 jig of NHS-LC-biotin (Pierce) per ml to a density of 5 x 109/ml and incubated at 37°C for 100 min. Before use in the binding and inhibition assays, the freshly biotinylated bacteria were washed extensively with saline. Isolated type 1 fimbriae (5 mg/ml) were incubated in bicarbonate buffer with NHS-LC-biotin (200 ,ug/ml) at 4°C for 2 h, washed with saline by using a Centricon microconcentrator, and stored at -40°C until use. Binding assay. Microtiter plates (Dynatech) were individually coated with a dilution row of each glycoprotein ranging from 2 to 500 nmol/liter. NCA-50, ribonuclease B, fibronectin, TEX-75, ovalbumin, or fetuin were resuspended in 70 p.1 of coating buffer (35 mM NaHCO3, 15 mM Na2CO3, 0.02% NaN3 [wt/vol], pH 9.6) and wells were coated by incubation at 37°C for 2 h. Nonspecific binding sites were blocked with 200 ,ul of 1% bovine serum albumin (BSA) (wt/vol) in saline at 37°C for 1 h. Biotinylated bacteria (5 x 109/ml) or biotinylated fimbriae (20 pg/ml) were resuspended in saline and incubated at 37°C for 1 h in a volume of 50 p.l per well. The plate was incubated with 100 pu1 per well of streptavidinalkaline phosphatase conjugate diluted 1:4,000 in TBS at 37°C for 1 h, and enzyme activity was visualized after the addition of 100 pl. of p-nitrophenylphosphate (1 mg/ml containing 10% [vol/vol] ethanolamine, pH 9.8) per well. Color development was recorded after 1 h at 37°C by monitoring the A405 values on a Titertek Multiscan ELISA reader. Between each incubation step, the plate was washed three times with 0.05% Tween 20 in PBS (vol/vol). Various negative controls, prepared by coating the wells with coating buffer only, adding unbiotinylated bacteria, not using bacteria, or using E. coli HB101 instead of E. coli 38, were used. Inhibition assay. Inhibition studies were carried out as described above for the binding assay using different sugars (MMan, MGlu, and MGal) on NCA-50-coated microtiter plates (4 ,ug/ml). The biotinylated bacteria or type 1 fimbriae were preincubated with various sugar concentrations (0.005 to 100 mM) for 30 min at room temperature before being added to the wells. Controls were prepared by incubating E. coli 38 in the absence of sugars, using E. coli HB101, and using wells coated with coating buffer only. Electron microscopy. Samples of bacteria and type 1 fimbriae were applied to Formvar-coated and glow-dis-

VOL. 59, 1991

BINDING OF NCA TO E. COLI EXPRESSING TYPE 1 FIMBRIAE

charged copper grids, negatively stained with 1% phosphotungstic acid for 30 s and examined with a Philips transmission electron microscope model CM 10 operating at 80 kV. Binding of NCA-50 to E. coli 38 was carried out in a volume of 20 (0.5 mg of NCA-50 per ml) for 4 h. Unbound NCA-50 was removed by centrifugation, and the bacteria were washed with saline, applied to a grid, and stained as described above. Incubation with ovalbumin or incubation of E. coli HB101 (Fim-) with NCA-50 or ovalbumin served as negative controls. For immunolabeling, the procedure described above was modified to allow the incubation of NCA-50 directly on grids coated with bacteria. For these studies, 10 pI of a bacterial suspension (109/ml) was adsorbed to a grid for 2 min at room temperature, the excess liquid was removed, and the grid was placed face down on a drop of 3% BSA in saline (wt/vol) for 30 min to reduce nonspecific binding. All of the following incubation steps were performed in saline with 3% BSA. The BSA-blocked grids were placed on a drop of NCA-50 (0.5 mg/ml) for 4 h at room temperature. After three washes with saline, the grids were incubated with T84.1E3 (1 p.g/ml) for 30 min, washed again, incubated with a 1:20 dilution of goat anti-mouse antibody conjugated with 5-nm-diameter gold particles (Janssen Life Sciences Products) for 30 min, washed, and stained as described above. Controls were performed under the conditions as described above but without the incubation with either T84.1E3 or NCA-50. Fluorescence-activated cell sorting (FACS) analysis of fMLP-activated PMNs. Polymorphonuclear leukocytes (PMNs) were isolated from whole blood by centrifugation through Percoll (19) and washed repeatedly with PBS. The purified PMNs were resuspended in PBS with 3% BSA (wt/vol) to a concentration of 2 x 106 cells per ml and stimulated with formylmethionyl-leucyl-phenylalanine (fMLP) at a concentration of 10-5 M. Following incubation periods of 0, 5, 10, 20, 40, 60 and 80 min, the stimulated cells were washed twice in PBS and resuspended in 50 RI of labeling buffer (PBS containing 3% BSA [wt/vol], 0.03% NaN3 [wt/ vol], 5 ,ug of cytochalasin B [wt/vol]) at 4°C. The PMNs were then incubated with F(ab')2 fragments of the NCA-reactive monoclonal antibody T84.1E3 (2 ,ug/50 ,lI) for 1 h at 4°C. After being washed, the cells were resuspended in 100 p.l of the labeling buffer and again incubated for 1 h at 4°C with a goat anti-mouse fluorescein isothiocyanate (FITC)-labeled F(ab')2 fragment (Jackson Immunoresearch) diluted 1:100 with labeling buffer. The labeled cells were then washed several times with PBS and analyzed with a Becton Dickinson FACS IV flow cytometer. Negative controls included elimination of the primary T84.1E3, F(ab')2 fragment or substitution of the secondary antibody with a FITC-labeled mouse anti-human F(ab')2 fragment (Jackson Immunoresearch). RESULTS

Carbohydrate differences in NCA glycoforms. Two purified glycoforms of NCA, NCA-50 (50 kDa) and TEX-75 (75 kDa), were digested with endo H in order to compare their respective complement of high-mannose oligosaccharides. This enzyme treatment was chosen by considering the carbohydrate specificity of type 1 fimbriae. Treatment with endo H decreased the apparent molecular weight of NCA-50 to approximately 40,000 but had little effect on the molecular weight of TEX-75 (Fig. 1). Taking into account that the molecular masses of fully deglycosylated NCA-50 and TEX-75 are both 35 kDa (21), the decrease in molecular size

Mr( 103)

a

b

c

2487

d

92 69-

i

4630 -

FIG. 1. Western blot analysis of NCA-50 (lanes a and b) and TEX-75 (lanes c and d) before (lanes a and c) and after (lanes b and d) treatment with endo H. The samples were electrophoresed (10% acrylamide gel), blotted onto a nitrocellulose filter, and detected as described in the text. The Mr values of the standard proteins used (Amersham rainbow markers) are indicated to the left of the gel.

after endo H treatment suggests that >50% of the glycosylated side chains of NCA-50 are high in mannose, whereas TEX-75 contains none or few high-mannose side chains. The carbohydrate compositions of NCA-50 and TEX-75 are presented in Table 1. NCA-50 contains considerably more mannose (a saccharide found in all N-linked oligosaccharides) and less galactose (indicative of complex oligosaccharides) than TEX-75, results consistent with the presence of more high-mannose structures in NCA-50. The values for NCA-50 are in close agreement with those values obtained for granulocyte NCA-50 isolated from normal spleens (11). Bacterial binding to NCA-50 and TEX-75 in solution. NCA-50 exhibited strong binding properties to a strain of E. coli (designated strain 38) isolated from a patient with a urinary tract infection. This strain was found to express type 1 fimbriae by a yeast agglutination assay and by detection with anti-type 1 polyclonal antibodies (data not shown). Incubation of E. coli 38 with 2 puM NCA-50 resulted in visible agglutination of the bacterial suspension, suggesting the presence of multiple binding sites on NCA-50 (Fig. 2A). In comparison, incubation of E. coli 38 with TEX-75 failed to result in any detectable agglutination (Fig. 2B). For controls, we also incubated E. coli 38 with other glycoproteins such as ovalbumin, thyroglobulin, or ribonuclease B that contain high-mannose oligosaccharide chains, and fetuin which has complex chains. None of these controls agglutinated this TABLE 1. Carbohydrate composition of NCA-50 and TEX-75

Composition" (mol%) (range)

Sugar

Sialic acid Fucose Mannose Galactose N-Acetylglucosamine

N-Acetylgalactosamine

NCA

3.1 11.3 46.0 14.0 34.7 1.0

(2.8-3.5) (10.6-11.4) (45.3-46.8) (14.0-14.2) (33.7-35.7) (0.9-1.0)

TEX

2.5 (0.7-4.3) 11.9 (11.5-12.2) 23.5 (23.5-23.6) 21.7 (20.7-22.6)

38.5 (37.8-39.1) 1.8 (1.6-2.0)

a Mean value for four determinations. Hydrolysis and analysis conditions are provided in the text. For comparison, the mole percentage values previously reported (11) for granulocyte NCA from normal spleens are as follows: 2.5 to 5.8 for sialic acid, 8.3 to 12.2 for fucose, 37.8 to 50.6 for mannose, 12.1 to 12.8 for galactose, 22.1 to 31.4 for N-acetylglucosamine, and 0 for N-acetylgalactosamine.

2488

SAUTER ET AL.

*:j§,.';$-#B3_tr.gi;¢:v >T',e.

.oa.,s# . -.? .R B;

A

.

v

tv +

q4 .... .: 4tt

;